internal consistency of teachers' decision-making: What factors are taken into account in deciding for ..... This would be an illustration of the limits of .... Miller, M., Frazier, C.C. & Richey, D.D. (1980) Student Nonpromotion and. Teacher Attitude ...
European Educational Research Journal, Volume 2 Number 2, 2003
The Role of Teacher Attitudes and Judgement in Decision-making: the case of grade retention PATRICK BONVIN University of Fribourg, Switzerland
ABSTRACT This article presents a study of grade retention as it ‘naturally’ occurs, focusing on teachers’ role in the process leading to retention or promotion for comparable peers. Achievement and IQ were measured prior to the retention decision. A total of 4248 second grade students initially participated in the study, of which 2.3% were retained. Eighty-three retained students were involved in the study, and matched with 83 low-achieving peers. Measures at the teacher’s level included attitude towards the efficiency of retention, towards criteria that ought to be taken into account in the decision and general attribution of children’s difficulties in school. Teacher decision-making shows internal consistency but seems biased when objective criteria are examined. The probability of a child being retained was substantially influenced by teachers’ attitude towards the efficiency of retention as well as by their evaluations of developmental maturity, intellectual potential and their achievement expectancies in language.
Introduction The belief that students learn better in homogeneous groups is still widely reflected in the structure and organisation of most elementary school systems (Oakes & Heckman, 1995). Practices such as segregation from regular classes and keeping a student back in the same grade to repeat the curricula (grade retention) are persistent witnesses of that meta-conception of learning (Oakes, 2001), although they have both been shown to have detrimental effects on children’s achievement and adjustment to school (see, respectively, Holmes, 1989; Jimerson, 2001; and Osborne et al, 1991; Haeberlin et al, 1999; Bless, 2002). This meta-conception of school learning is related to further negative
277
Patrick Bonvin
consequences when one considers the interrelationships between the two practices. There is, for instance, some evidence that retention often precedes segregation and is applied prior to the diagnosis of a learning difficulty (McLeskey & Grizzle, 1992; Barnett et al, 1996), with the adverse consequence of delaying the child’s referral for remedial services. More than half of the students diagnosed with a learning difficulty had already been retained at the time of diagnosis. In fact, it seems that children displaying both a learning difficulty and behavioural problems benefit from accelerated referral, and thus a better prognosis, than children with isolated learning difficulties (McLeskey & Grizzle, 1992). Dropping out, as a deferred form of exclusion from the regular class, is also predicted by earlier retention (Roderick, 1994; Jimerson et al, 2002). These problems are relevant for most European educational systems (during the school year 2000/2001 only Iceland, the United Kingdom, Sweden and Norway did not apply grade retention, cf. Eurydice, 2003). They will become more and more crucial as the pressure for accountability and efficiency increases (e.g. Allington & Mc Gill-Franzen, 1992). The study presented in this article stems from a research programme on inclusion initiated in the mid-1980s in Switzerland in order to evaluate the efficiency of different schooling systems in dealing with special needs (IntSep: see Haeberlin et al, 1999; Kronig et al, 2000; Bless, 2002). Parallel analysis of school statistics revealed that although segregation was rising, retention rates did not decrease: when more of the ‘lowest’ achievers are removed from the regular classroom, the threshold for retention increases (Bless & Kronig, 1999). This keeps raising the selectivity of the system, leaving more and more children behind. As a further study of the ‘inclusion ability’ (Integrationsfähigkeit) of the Swiss schooling system, the latest phase of the programme (IntSep-R, 2000-2003) is thus dedicated to the study of grade retention. Its main research questions are concerned with (a) the determinants of retention, (b) the effects of grade retention on the children’s academic achievement, and (c) the social and emotional consequences of grade retention. The present article focuses on the first of those areas of investigation and seeks a better understanding of why students are selected for retention, as what can often be considered as the first step in a process leading to exclusion from the regular curriculum. Framework It seems that grade retention will be applied whether efficient or not. A basic requirement would then be that the criteria for retention be explicit and just. We are lacking research on how students become eligible for retention. Of course, the achievement of retained students is below the average achievement of the population to which they belong (Cadigan et al, 1988; Alexander et al, 1994). This is true also where IQ is concerned (Safer, 1986). In addition, classical demographic risk factors for difficulties in school, such as socio-economic status and ethnicity/nationality, appear to influence the 278
THE ROLE OFTEACHER ATTITUDES AND JUDGEMENT
process leading to retention (Hauser et al, 2001). Another relevant demographic characteristic is the pupils’ sex: boys are retained more frequently than girls (Hauser, 1999). In this respect, the process leading to retention looks similar to the one leading to segregation, boys being more liable than girls to be affected by such decisions (Bennett et al, 1993). These factors emerge in studies where the children’s initial level of achievement was not controlled. They give information about ‘who’ is at risk for repeating a grade, but do not answer the question ‘why’, among the at-risk children, a child will repeat a grade while a comparable peer will be promoted. The study of comparable students is necessary to answer the latter question. As stated by Cadigan et al (1988), the study of children at risk to be retained can be likened to the study of ‘outliers’ (p. 73). Studies of the predictors of retention comparing matched samples of retained and promoted students are scarce. In a study of retention after the first primary grade, Cadigan et al (1988) report significant predictors as being reading grades given by teachers in the first quarter of the school year, kindergarten teachers’ observations of children’s cognitive, motor and affective development, and actual teachers’ evaluations of children’s popularity. Predictors were entered in a four-block logistic regression (with retention vs. promotion as a dichotomous dependent variable) representing teacher characteristics, parents’ characteristics, children’s achievement (grades) and personal characteristics. The only blocks to yield significant results were those associated with children’s characteristics and achievement assessed by teachers. This is striking considering that the most powerful predictors in the model are the grades given by teachers after only two months into the school year, in a subject for which the achievement of retained students does not differ from that of promoted students at the end of the year (before the retention). The authors conclude, ‘Our analysis suggests that … the decision to promote or not promote was influenced by only a few factors, which related mainly to teacher judgments of the child’ (Cadigan et al., 1988, p. 87). These results were replicated in the follow-up of the study (Dauber et al, 1993), although samples were not matched. The general results nevertheless point towards the important role played by teacher evaluations and grades even in comparison with students’ performance on standardised achievement tests. The teacher’s role in the process leading to grade retention has also been studied in the context of research explaining the persistency of the practice despite negative empirical evidence or official policy statements (cf. AndersonLevitt et al, 1991). An obvious reason for teachers to apply grade retention is that they believe it is effective. This belief is indeed true for the majority of teachers in most studies, although answers show high variability (Miller et al, 1980; Byrnes & Yamamoto, 1986; Pini, 1991; Tomchin & Impara, 1992). The most important criteria for retention mentioned by teachers in Tomchin & Impara’s (1992) study are low achievement, low ability, immaturity and lack of effort (see also Pini, 1991). Here too, high variability is to be found (teachers opt for retention for different reasons). The particular weight put on 279
Patrick Bonvin
achievement and immaturity as reasons to recommend retention (both come out strongly from the two studies mentioned above) is probably explainable by the very nature of the measure. The idea behind retention is that by spending an extra year in the same grade the student gets the opportunity and time to ‘catch up’ by repeating the programme, or that the extra year allows the child to gain in maturity. Unfortunately, we are lacking evidence asserting that beliefs actually result in different practices (for example, that teachers who believe retention is an efficient practice tend to apply it more than those who do not). A simpler indicator of the importance of teachers’ role in deciding for retention or promotion is to consider the rules of practice in the school system studied; in our case (Switzerland), the decision lies predominantly in the teacher’s hands (Information Documentation Education Suisse [IDES], 2002). Conclusions of research on teacher attitudes towards retention usually focus on the variability of individual teachers’ practices, although it may relate only to the more detailed features of the practice, for example, the criteria on which the decision is based, rather than to the general utility of the measure, for which teachers exhibit convergence with popular thinking about retention (Miller et al, 1980). This conclusion is similar to that of Pini (1991), who observes variability in teacher attitudes but not in the global belief about the pedagogical value of grade retention. Aims of the Study The general aim of the study presented in the following sections is to document the teacher’s role in the decision to promote or retain a child by examining the effects of attitudes towards retention and evaluations of the student on the probability of retention beyond children’s objective characteristics. What we want to describe, as a first set of questions, are teachers’ attitudes towards retention: its general efficiency and what factors should be considered in the decision (more details are available in the methodology section). A supplementary question is added in reference to work on interpersonal attribution processes in education (Weiner, 2000): what are teachers’ general attributions about the causes of school failure? It has indeed been demonstrated that the attributions teachers make about successes or failures affect their behaviour towards students (Weiner, 2000; Reyna & Weiner, 2001). The dimensions of controllability and locus of control influence this process in particular by affecting teachers’ feelings and behaviours. When failure is attributed to uncontrollable factors, teacher attitudes towards the pupil will be more empathic and lead to helping behaviours, whereas, if failure is attributed to factors under the pupil’s control, anger will predominate and the teacher will be less likely to provide help to the student (Weiner, 2000; Georgiou et al, 2002). These mechanisms are relevant to our approach, since the type of causal attribution favoured by teachers seems to be a potential facilitator of rejection and exclusion (Reyna & Weiner, 2001). High attribution 280
THE ROLE OFTEACHER ATTITUDES AND JUDGEMENT
of responsibility and an inclination for dispositional attribution will then be considered as hypothetically linked with heightened use of retention. The second set of questions links the results of the above-mentioned descriptive questions to the actual practice of retention. A preliminary step will be to investigate the relationships between attitudes, particularly those between criteria for retention and general causes of failure on the one hand, and attitudes towards the efficiency of grade retention on the other. According to theories about the attitude–behaviour relation (see Eagly & Chaiken, 1998), the attitude the most proximal to the behaviour (usually, behavioural intention; cf. Ajzen & Fishbein, 1980) is the most likely to have an effect on the behaviour’s probability of occurring. This is consistent with propositions by action theorists (e.g. Orton, 1996), according to which the basic relationship between attitudes and behaviour can be conceptualised in line with the classical practical syllogism. The attitude, or belief, fills the gap between the goals pursued and the action chosen in order to reach them: ‘I want to resolve such and such a problem [goal], I think/believe such and such a method is efficient in dealing with this particular problem [belief and/or knowledge], so I apply such a method in this case [action]’. This model is presented as an idealised process of reasoning about action (Orton, 1997). To conclude, it is plausible that teachers’ attitudes towards the efficiency of retention may mediate the relation between the other two sets of attitudes (criteria and causes) and the practice, as a sort of necessary path to retention. The remaining questions refer to teacher evaluations of students and their relationship to the practice of retention: how is the decision influenced by specific teacher judgements of student characteristics and achievement expectations? This question in turn implies an additional question about the internal consistency of teachers’ decision-making: What factors are taken into account in deciding for promotion and retention? Methodology Participants and Research Plan To investigate these questions, the following research plan was embedded in a larger project aiming at the study of retention as it ‘naturally’ occurs, focusing on the effectiveness of the measure (cf. Schuepbach, 2002). A total of 4248 second grade students (234 classes) took part in a first measurement point and at the end of the school year. Regional and local educational authorities were involved in the sampling procedure. Teachers were sampled on the basis of opportunity sampling. The number of classes for each region (canton) was defined according to the size of the population. Other sampling principles involved respecting town/land ratios as well as representative ratios of minority students within regions. Sexes are equally represented (49.4% girls and 50.6% boys). Ages range from 6 years and 9 months (6:9) to 11:3 (M = 8:8). After teachers communicated the promotion–retention decision, retained
281
Patrick Bonvin
students (n = 99) were paired with promoted students comparable on measures of IQ, achievement in mathematics and language, nationality and gender. After the pairing procedure was completed, 83 retained–promoted pairs with complete data could be formed. The quality of pairs was checked by between-groups chi-square tests and t-tests on the following variables: sex, nationality, mother tongue, level of mastery of instruction language, age, IQ, achievement in mathematics and language. The two groups did not differ on these measures (no test reached significance). Follow-up of students in subsequent years is not included in this research as the focus is on the determinants of retention. Comparison of retained students with the whole sample yields results compatible with expectations derived from classical risk factors for low achievement in school: more boys (55% vs. 50.6%) and more minority students (38.6% vs. 22.3%) are to be found in the retained sample. IQ and achievement results are also significantly lower. Some 91.8% of teachers are women. Average age is 39 (minimum = 20, maximum = 62; professional experience: M = 15 years). Teachers give information about themselves, their attitudes towards retention and school failure, their schools and classrooms, and student demographic data at their enrolment in the study. They also evaluate specifically each student on behavioural variables and social integration, and give an estimate of students’ performance on the achievement tests. Data can then be available on three levels: teacher attitudes (for example, how important is the factor achievement in mathematics in deciding to retain a student?), teacher evaluation of a given child’s expected performance on the mathematics test, and the objective level of performance of that child on the test. Instruments Teacher Questionnaire (TQ). The questionnaire was constructed in order to gather data related to teachers’ personal and professional characteristics as well as to their assessment of criteria for retention, their general attributional tendency regarding school failure, and their attitude towards the efficiency of grade retention. Questions about the weight of criteria for retention (18 items) and estimates of the importance of various causes of failure (24 items) were presented as four-level Likert scales (ranging from ‘not at all important’ to ‘very important’). Evaluation of teacher attitude towards the efficiency of retention was also presented along a four-level Likert scale (‘not at all efficient’ to ‘very efficient’). Items presented in the ‘criteria for retention’ and ‘causes of failure’ lists were chosen according to literature specific to retention (e.g. Tomchin & Impara, 1992), to teacher-attribution literature (e.g. Georgiou et al, 2002), psycho-educational decision-making literature (e.g. Andrews et al, 1997) or general literature about the factors of success in school (including structural factors; e.g. Wang et al, 1993).
282
THE ROLE OFTEACHER ATTITUDES AND JUDGEMENT
Teacher–Student Questionnaire (TSQ). The instrument collecting teachers’ evaluations and information about students was constructed as a series of lists of pupils, including scales and coding information on the same sheet. The following information was provided: nationality, mother tongue, mastery of the language of instruction, special interventions within or outside the school, social integration, intellectual potential, maturity, motivation, aggression, disruptive behaviours, and expectancies regarding performance on the mathematics and language tests. A space for personalised remarks was provided. This system allows the teacher to enter only data that make sense (e.g. no entry in the nationality field if the pupil is Swiss), without having to consult a separate coding manual. Children’s Objective Characteristics IQ. The CFT-1 (Culture-fair intelligence test [Cattell et al, 1997]) was used. It is adapted to the age group and to group testing, not sensitive to prior learning (especially in the verbal area), and allows fine discrimination between middle and low intelligence (this was especially important, as the project was aimed at a population of low achievers). Achievement tests. The Swiss schooling system is particular where subject matter curricula are concerned. Curricula are often defined regionally and some cantons have their own curriculum. The procedure for test construction involved the gathering of all available curricula, a compilation of major objectives of mastery for the second grade, and, finally, the extraction of common elements. Initial versions of mathematics and language tests were constructed and pre-tested in both linguistic regions (French and German speaking). On that basis, items were selected on the basis of their relation to the total score (item–total correlation), their difficulty and their consistency. Analysis Teacher attitude towards the efficiency of retention, attitudes towards the criteria for the decision and towards the main causes of school failure were initially analysed descriptively on the item level. Item average scores were used as indicators of the respective importance attributed to each item. Factor analysis (PCA, Varimax rotation) was then applied to both groups of items (criteria for retention and causes of failure), which yielded the following factors: child/school-related (e.g. attention, work behaviour, etc.), child/personal (e.g. psychological problems, physical health), family (e.g. socio-economic status of family, support at home), achievement, and intelligence where criteria are concerned, to which one should add a system factor (selectivity, means), teaching (teacher competency, personality, differentiation) and a factor grouping children’s nationality and language, when one considers the general causes of failure in school. Results pertaining 283
Patrick Bonvin
to the relationships between attitudes and to the predictors of retention are presented according to scores on the above-mentioned factors. Relationships between factors (see Tables I and II) are analysed using multiple regression. In analyses predicting retention, the dependent variable is the retention vs. promotion dichotomy (for matched samples). Belonging to either group is predicted by logistic regression analysis. Results Attitudes Towards the Efficiency of Grade Retention In order to understand results, it is important to note that all analyses involving teacher attitudes comprise the whole sample (classes n = 234, students n = 4248). On the other hand, analyses predicting retention vs. promotion concern the subsample composed of retained students and comparable peers (n = 166, 83 pairs). Descriptive results pertaining to teachers’ attitude towards the efficiency of grade retention for dealing with low achievement in school show that teachers do not seem to generally favour grade retention. In fact, the ratio of positive to negative attitudes is nearly 50%. This corresponds to the high variability found in similar studies. Teachers do not seem to share a consensus on the usefulness of grade retention. Criteria for Retention Our next question deals with factors to be taken into account in the decision process. On average teachers evaluated the most important criterion in the decision to be developmental immaturity. This is followed by all achievement variables (mathematics, language and general), and by intellectual potential (aptitudes). These three factors correspond to those mentioned in previous research (Pini, 1991; Tomchin & Impara, 1992). Lack of effort, the fourth factor frequently mentioned as important, was not measured as such in this study. It partly fits into the concept of work behaviour, which is an important factor here. This factor is, however, preceded in our study by psychological problems. Causes of School Failure Psychological problems appear to be the most important factor for failure in school according to teachers. Developmental maturity is again mentioned as very important, as well as intelligence. Other child-related variables mentioned as important are attention, motivation and work behaviour. Achievement-related variables and teacher-related variables (competence, personality and differentiation) are rated in general as less important than those factors given above, but are still quite strong factors in school failure according to teachers. School system and family-related variables are usually among the least 284
THE ROLE OFTEACHER ATTITUDES AND JUDGEMENT
important factors, apart respectively from number of students in class (midrange) and family situation (among the most important factors). As with criteria for retention, child-related variables are those factors most frequently mentioned by teachers as very important causes of school failure. From an attributional perspective, it seems that the most important factors are shared between internal stable factors (psychological problems, maturity) and internal controllable causes (concentration, motivation, work behaviour). Teachers do not, however, seem to underestimate their own role, as teacher-related variables are on average mentioned as quite important. Causes that come out as less important are system–school factors, such as selectivity, and contextual– family factors such as lack of support at home or socio-economic status of the family, the latter being the least important factor. The Relationship between Attitudes towards Retention, Criteria and Attributions The next step in the analysis involves investigating relationships between attitudes towards criteria, general attributions for causes of school failure and attitude towards the efficiency of grade retention. Results are shown in Table I where the decision criteria are concerned and in Table II for general attribution of failure, according to the factorial structure of the instruments (see Analysis section). Factors
B
S. E.
Beta
t
Sig.
Child: school-related
.09
.07
.11
1.42
.16
Family
-.07
.07
-.08
-1.03
.30
Achievement
.16
.07
.15
2.31
.02
Child: personal
-.06
.08
-.06
-.74
.46
Intelligence
.06
.05
.08
1.18
.24
Dependant variable: teachers' attitude towards retention Table I : Weight of factors in the retention decision
The only clear effect is that teachers who value achievement most as a decision criterion tend to have more positive attitudes towards retention (β = 0.16; p